Comp strategy
Fidelity is a specification,
not an afterthought.
Most comp requests fail before the vendor reads them. Not because the brief was wrong, but because it never said what the comp had to do.
Most comp requests fail before the vendor reads them. Not because the vendor is wrong, not because the timeline is unrealistic, and not because the budget is too tight. Because the request didn’t specify what the comp needed to do.
Fidelity is the most consequential variable in a physical comp. It determines cost, timeline, and whether the sample can actually answer the question it was built to answer. And in most briefs, it goes unnamed.
That gap is where budget leaks and bad data are born.
The problem
Friction that traces back to one missing word.
Brand teams and comp vendors experience the same friction, described in different languages. The brand team says the comp came back wrong. The vendor says the brief was vague. Both are correct.
The typical comp request specifies structure, materials, and graphics. It describes the object. It doesn’t describe the decision. And fidelity, correctly understood, isn’t a property of the object. It’s a property of the decision the object exists to inform.
When fidelity goes unspecified, two things happen. Both are expensive.
Over-building → wasted spend
A production-matched comp gets ordered for a decision that a rough mock would have settled in two days for a fraction of the cost. No one calls this waste explicitly. It gets absorbed into the project budget and attributed to the category.
Under-building → wrong data at full production cost
A rough mock gets submitted as stimulus for shopper research. The research runs. Confident numbers land in a deck, travel to leadership, and shape decisions about tooling, inventory, and launch sequencing. The mock was wrong. The data was wrong. The launch answered the wrong question at full production cost.
Under-building doesn’t fail loudly. That’s what makes it dangerous. A slightly-off stimulus doesn’t produce obviously bad data. It produces plausible data. Credible data. Data that gets cited in a gate review. As Luke Cahill has observed: in-context research is only as honest as the stimulus it’s built on.
The comp is the lowest-risk place in the system to be wrong. The launch is not.
In-context research is only as honest as the stimulus it’s built on.Luke Cahill
The definition
What fidelity actually means.
Fidelity describes how closely a comp replicates the production artifact across the dimensions that matter for a specific decision.
Two comps built from the same dieline can have entirely different fidelity levels. A label wrap around a generic substrate is low fidelity. It communicates graphic intent and hierarchy. It tells you almost nothing about how the structure performs on shelf, how it feels in hand, or how it renders under fluorescent retail lighting.
A production-matched comp is built to replicate material, weight, finish, structural performance, and print output as closely as the comp process allows. It’s a test sample, not a presentation prop. It’s built to answer questions that only the physical, finished form can answer.
Neither is inherently better. The right fidelity level is the one that matches the decision being made. The problem is that most comp requests don’t say which decision that is.
The framework
Five decisions, five fidelity levels.
Mapping fidelity to decision type removes ambiguity from the brief. These five scenarios cover the majority of comp use cases in brand and product development. The level climbs with the stakes, from internal alignment up to pre-tooling validation.
Internal alignment and hierarchy checks
Fidelity: LowThe decision: does the design system read correctly? Are visual hierarchy, color relationships, and layout working as intended?
Fidelity required: low. A label wrap on a proxy substrate is sufficient. Print accuracy matters. Structural accuracy doesn’t. This comp is a communication tool, not a validation tool.
Leadership and budget reviews
Fidelity: Low to mediumThe decision: does this design concept merit continued investment? Can decision-makers see the intent clearly enough to approve the next phase?
Fidelity required: low to medium. The comp should look intentional and finished at a surface level. It doesn’t need to replicate structural performance or material weight. It needs to read as credible in a conference room setting.
Buyer meetings and line reviews
Fidelity: Medium to highThe decision: does the retailer see the product as shelf-ready, competitively positioned, and appropriate for their format?
Fidelity required: medium to high. Buyers handle product. They place it on shelf fixtures. They compare it to existing SKUs. A comp that doesn’t hold its structure, reflect its actual material weight, or replicate its finish will produce feedback that reflects the comp, not the product.
Consumer research and shopper studies
Fidelity: High to production-matchedThe decision: how does the target shopper respond to this product when encountered in a realistic purchase environment?
Fidelity required: high to production-matched. This is the category where under-building is most consequential and most common. In-context research methodology is designed to produce behavioral signal from realistic conditions. When the stimulus isn’t realistic, the signal it produces isn’t about the product. It’s about the comp. The research appears to run correctly. The outputs appear actionable. They’re not. A comp built for consumer research must replicate the materials, finish, weight, and structural performance of the production unit closely enough that the research participant responds to the product, not the prop.
Pre-tooling and pre-production validation
Fidelity: Production-matchedThe decision: does this structure, material, or format perform as required before tooling investment is committed?
Fidelity required: production-matched. There’s no useful version of this comp that isn’t built to production specification. The only question it exists to answer is whether the production unit will work. A lower-fidelity comp can’t answer that question. Ordering one for this purpose isn’t a cost-saving decision. It’s a deferred production error.
The solution
A before-and-after comp request.
The difference between a vague brief and a fidelity-specified brief isn’t length. It’s precision about the decision.
“We need a comp of the new 12oz format. Full graphics, final dieline. Needed by end of next week.”
This brief will generate a slow, vague quote. The vendor doesn’t know what decision the comp must inform. They don’t know what fidelity is appropriate. They’ll scope to a middle assumption, build to that, and deliver something that may or may not answer the question the brand team was actually asking.
“We need a comp of the new 12oz format to serve as stimulus for an in-context shopper study running in three weeks. The comp will be placed on shelf in a simulated retail environment and handled by research participants. It needs to replicate the production material weight, finish, and structural integrity of the final unit. Graphics are final. Please scope to production-matched fidelity.”
This brief generates a clear quote, a specific timeline, and a comp that can do the job it was ordered for. The difference is one sentence about the decision.
When you request a comp, give your comp partner three pieces of information. Those three answers define the required fidelity level. Everything else in the brief follows from there.
What decision this sample informs
Name the call the comp exists to support, not the object it depicts. The decision sets the bar for everything else.
Who handles it and under what conditions
An internal reviewer, a retail buyer, and a research participant each demand a different level of realism from the sample.
What would make the result untrustworthy
Define the failure. If a soft structure or an off finish would invalidate the outcome, the comp has to replicate it.
Fidelity isn’t a detail.It’s the brief.
Fidelity isn’t a detail to resolve after the brief goes out. Spec it to the decision, not the object, and every other variable in the comp process becomes easier to manage.
Before your next comp, name the decision. The fidelity level follows from there, and so does a faster quote, a cleaner build, and data you can trust.
Reach Bob directly · bob.jennings@3dcolor.com
Before your next comp, spec fidelity to the decision
Every comp, calibrated to the decision it has to inform.
3D Color builds production-matched packaging comps, prototypes, and sales samples for CPG brands, each one calibrated to the decision the sample has to inform. When fidelity is specified up front, the quote is faster, the build is right, and the data holds up.
Bob Jennings, CEO of 3D Color
Comps, prototypes, and sales samples are 100 percent of the business. If your team is briefing a comp and hasn’t yet named the decision it has to inform, he’s the person to reach.
Reach Bob directly